Learning Strategy Based on Deep Knowledge Tracing

Network-based classroom is intended to spread courses through the Internet, so that people can share educational resources through the Internet. The existing network-based classroom can collect students' video images, body behavior and other information, track students' learning status, including whether they listen carefully and answer questions actively. However, it ignores the real-time tracking of students' mastery of knowledge concepts and can not provide personalized learning strategy. In order to solve this problem, we design an learning strategy model based on deep knowledge tracking. The model is divided into two steps. Firstly, the model analyzes the sequence of students' exercises through deep learning algorithm, and predicts students' mastery of knowledge concepts. Then, according to the prediction results and students' learning situation, C4.5 algorithm is used to generate learning strategy decision tree to provide students with personalized learning strategy. The model is tested on two real datasets. The results show that the model has a good performance in predicting students' mastery of knowledge concepts and providing learning strategies.